Large-Scale Practice of Persistent Memory in Alibaba Cloud

The presentation discusses the benefits and potential use cases of personal memory devices in software architecture.

• Personal memory devices can provide a larger radius instance for database services.
• Using personal memory achieves 90% performance and 70% cost compared to DRAM.
• Personal memory devices can be used as high-performance local storage or memory cache.
• The field of memory pool can benefit from personal memory devices as chip memory sources.
• Efforts are being made to enhance device performance and implement memory dynamic provision in cognitive environments.

In one scenario, personal memory devices can be used as a memory cache to catch breaking data and split up applications. This allows the system to move memory data from offline applications to personal memory when the application begins to stop. When the offline application collapses, the system can impose the main data from host memory to DRAM immediately.

Persistent memory allows programs to access data as memory, directly byte-addressable, while the contents are non-volatile, preserved across power cycles. Alibaba have millions of databases and memory prefered applications which need massive of memory resource and make a huge cost every year. PMEM device provides the high performance and lower price which have been widely used in Alibaba Cloud. We have developed a combined system focus on PMEM resource optimization and capacity scheduler, which is widely used in our Kubernetes platforms. With the system, PMEM device is used as kuberentes volume object, and can be configed in different types: kmem, quotapath, lvm, direct. Also we optimize the scheduler to implement that PMEM device has best match with numa node.